A novel, highly selective, and general methodology for the synthesis of 1,5-diene-containing oligoisoprenoids of all possible geometrical combinations exemplified by an iterative and convergent synthesis of coenzyme Q10

Article abstract of DOI:10.1021/ol010263d

Chemical equation presented A truly general, versatile, and highly regio- and stereoselective methodology for the synthesis of terpenoids containing 1,5-diene units of E and/or Z geometry critically involves Pd-catalyzed homoallyl- and homopropargyl-alkenyl coupling and Zr-catalyzed carboalumination of alkynes. By using this methodology, coenzyme Q₁₀, (E,Z,E)-geranylgeranoil, and other natural or unnatural compounds have been synthesized efficiently.

Full text of DOI:10.1021/ol010263d

ORGANIC
LETTERS
2002
Vol. 4, No. 2
261-264
A Novel, Highly Selective, and General
Methodology for the Synthesis of
1,5-Diene-Containing Oligoisoprenoids
of All Possible Geometrical
Combinations Exemplified by an
Iterative and Convergent Synthesis of
Coenzyme Q
10
Ei-ichi Negishi,* Show-Yee Liou, Caiding Xu, and Shouquan Huo
Herbert C. Brown Laboratories of Chemistry, Purdue UniVersity,
West Lafayette, Indiana 47907-1393
negishi@purdue.edu
Received November 7, 2001
ABSTRACT
A truly general, versatile, and highly regio- and stereoselective methodology for the synthesis of terpenoids containing 1,5-diene units of E
and/or Z geometry critically involves Pd-catalyzed homoallyl- and homopropargyl-alkenyl coupling and Zr-catalyzed carboalumination of alkynes.
By using this methodology, coenzyme Q , (E,Z,E)-geranylgeranoil, and other natural or unnatural compounds have been synthesized efficiently.
10
We report herein an unprecedentedly efficient, highly selec-
tive, and general methodology for the synthesis of stereo-
and regiodefined isoprenoids containing 1,5-diene moieties
of all conceivable geometrical combinations via one-pot
chain elongation by one E or Z C5 isoprene unit predomi-
nantly in the “tail-to-head” (T-to-H) direction.1,2 The devel-
opment of the methodology was critically dependent on
several reactions and procedures we have previously dis-
covered and developed. First, screening of several candidates
led to the selection of (E)- and (Z)-1,4-diiodo-2-methyl-1-
butenes (1a and 1b) as difunctional (E)- and (Z)-isoprene
synthons permitting one-pot chain elongation with incorpora-
tion of an either (E)- or (Z)-isoprene equivalent that can be
repeated twice or more, while maintaining a very high
stereoselectivity level (vide infra). Throughout this study,
no stereoisomers were formed and/or detected even after
(2) The “head-to-tail” (H-to-T) and “tail-to-head” (T-to-H) directions in
isoprenoids are defined as shown below, for the sake of convenience.
(1) This work was orally presented in part as a short invited lecture at
the 10th IUPAC OMCOS Symposium, Versailles, France, July 18-22,
1999, and its abstract has been published as part of the symposium
proceedings: Negishi E.; Liou, S. Y.; Xu, C.; Huo, S. Polyhedron 2000,
19, 591.
10.1021/ol010263d CCC: $22.00 © 2002 American Chemical Society
Published on Web 01/04/2002

many synthetic operations as long as the E synthon (1a) was
used. We therefore judge that the stereoisomeric purity of
1a that is readily preparable from 3-butyn-1-ol in two steps3,4
in 77% overall yield must be infinitesimally close to 100%
(Scheme 1). The Z isomer of 1a, i.e., 1b, has recently become
now be achieved for efficient, clean, and iterative construc-
tion of oligoisoprenoids. Although we have previously
reported the treatment of homoallyl and homopropargyl
halides containing Br or I with Mg in the presence of zinc
halides as a satisfactory procedure,6 it is less clean, producing
some dimeric and other byproducts, and less dependable in
highly demanding cases, such as those discussed herein. The
difference is further accentuated in iterative synthetic opera-
tions without isolation of intermediates.
Scheme 1
Fourth, the use of 1 requires terminations at both ends of
an isoprenoid chain termed “head capping” and “tail
modification” for the sake of convenience. Head capping
merely requires the Pd-catalyzed homoallyl-alkenyl coupling
discussed above with an appropriate 2-methyl-1-alkenyl
iodide, while modification of terminal alkynes at the tail to
produce (E)- and (Z)-3-methyl-2-alken-1-ols is most conve-
niently achieved by using our one-pot Zr-catalyzed methyl-
alumination-ate complexation-hydroxymethylation11 and
Sato’s two-step alkyne hydroxymethylation-Ti-catalyzed
syn-hydromagnesation-methylation,12 respectively.
The potential synthetic utility of the protocol incorporating
those features discussed above is indicated by an efficient
and selective iterative and convergent synthesis of coenzyme
Q10 (2)13 from 1a, 4-iodo-1-trimethylsilyl-1-butyne (3), (E)-
1-iodo-2,6-dimethyl-1,5-heptadiene (4), and 2,3-dimethoxy-
4-chloromethyl-5-methyl-1,4-benzoquinone (5)14 in nine
steps in 26% overall yield (Scheme 2). No stereoisomeric
separation was attempted in the synthesis of 98% (all-E)-2.
It should also be noted that all nine stereodefined C5 isoprene
units were linked and incorporated into coenzyme Q10 (2)
only in nine synthetic operations.
The reaction of alkenylalanes with allyl or benzyl chlorides
can be achieved by using either Ni or Pd catalysts, as we
originally reported in 1981,15 but the Ni-catalyzed reaction
revealed a minor tendency to undergo double bond isomer-
ization of the cross-coupled products in some cases. A more
recent report14 claimed the superiority of Ni catalysts over
Pd catalysts in terms of reaction rate and product yield, and
we in fact completed our synthesis of coenzyme Q10 using
a Ni catalyst, which proved to be highly satisfactory. Further
experiments summarized in Scheme 3, however, suggest that
Pd catalysts are also very satisfactory in cases where
chloromethylquinones are used as the electrophilic cross
coupling partner.
accessible via thermal rearrangement of the product obtained
by Zr-catalyzed carboalumination of 3-butyn-1-ol,5 although
the long reaction time for thermal isomerization and the 98%
isomeric purity level leave some room for improvement
(Scheme 1).
Second, the Pd-catalyzed homoallyl-alkenyl and homopro-
pargyl-alkenyl cross coupling of 1 with homoallylzinc and
homopropargylzinc6 derivatives developed by us can proceed
selectively and in high yields without a sign of any
isomerization through the use of Cl2Pd(dppf)7 as a catalyst.
3
Particularly noteworthy is that the Csp -bound homoallylic
or homopropargylic I or Br does not compete to detectable
3
2
extents with the Csp -bound I or Br, as reported first by us,
permitting clean and selective elongation with 1a or 1b in
the T-to-H direction.
Third, lithiation of primary alkyl iodides via Li-I ex-
change reported by us8 and Bailey9 in 1990 leads to the
formation of the corresponding alkyllithiums generally in
almost quantitative yields with little or no sign of any side
reactions even in those cases where the alkyl groups are
homoallyl, homopropargyl, and homobenzyl.10 Coupled with
another essentially quantitative Li-to-Zn transmetalation, an
exceedingly clean and highly dependable generation of the
requisite homoallyl- and homopropargylzinc derivatives can
(11) Okukado, N.; Negishi, E. Tetrahedron Lett. 1978, 2357.
(12) Sato, F.; Ishikawa, H.; Watanabe, H.; Miyake, T.; Sato, M. J. Chem.
Soc., Chem. Commun. 1981, 718.
(3) Rand, C. L.; Van Horn, D. E.; Moore, M. W.; Negishi, E. J. Org.
Chem. 1981, 46, 4093.
(4) For seminal papers on the Zr-catalyzed carboalumination, see (a) Van
Horn, D. E.; Negishi, E. J. Am. Chem. Soc. 1978, 100, 2252. (b) Negishi,
E.; Van Horn, D. E.; Yoshida, T. J. Am Chem. Soc. 1985, 107, 6639. (c)
Negishi, E. Pure Appl. Chem. 1981, 53, 2333.
(5) Ma, S.; Negishi, E. J. Org. Chem. 1997, 62, 784.
(6) (a) Negishi, E.; Valente, L. F.; Kobayashi, M. J. Am. Chem. Soc.
1980, 102, 3298. (b) Kobayashi, M.; Negishi, E. J. Org. Chem. 1980, 45,
5223.
(7) Hayashi, T.; Konishi, K.; Kumada, M. Tetrahedron Lett. 1979, 1871.
(8) Negishi, E.; Swanson, D. R.; Rousset, C. J. J. Org. Chem. 1990, 55,
5406.
(13) (a) Littaru, G. P.; Ho, L.; Folkers, K. Int. J. Vitam. Nutr. Res. 1972,
42, 291. (b) Blizmakov, E. G.; Hunt, G. L. The Miracle Nutrient Coenzyme
Q10, Bantom Books, New York, 1987. For representative approaches to
the synthesis of these targets, see (c) Eren, D.; Keinan, E. J. Am. Chem.
Soc. 1988, 110, 4356. (d) Ru¨ttimann, A.; Lorenz, P. HelV. Chim. Acta 1990,
73, 790. (e) Yanagisawa, A.; Nomura, N.; Noritake, Y.; Yamamoto, H.
Synthesis 1991, 1130. (f) Lipshultz, B. H.; Bulow, G.; Fernandez-Lazaro,
F.; Kim, S.-K.; Lowe, R.; Mollard, P.; Stevens, K. L. J. Am. Chem. Soc.
1999, 121, 11664 and references therein.
(14) (a) Lipshutz, B. H.; Bulow, G.; Lowe, R.; Stevens, K. L. Tetrahedron
1996, 52, 7265. (b) Lipshutz, B. H.; Kim, S.-K.; Mollard, P.; Stevens, K.
L. Tetrahedron 1998, 54, 1241.
(15) (a) Matsushita, H.; Negishi, E. J. Am. Chem. Soc. 1981, 103, 2882.
(b) Negishi, E.; Matsushita, H.; Okukado, N. Tetrahedron Lett. 1981, 22,
2715. (c) Negishi, E.; Chatterjee, S.; Matsushita, H. Tetrahedron Lett. 1981,
22, 3737.
(9) Bailey, W. F.; Punzalan, E. R. J. Org. Chem. 1990, 55, 5404.
(10) The use of alkyl bromides and secondary alkyl iodides leads to
complex and/or unsatisfactory products.
262
Org. Lett., Vol. 4, No. 2, 2002

Scheme 2
In addition to high efficiency and selectivity, the general
by one isoprene unit. Furthermore, it is as such not applicable
to the synthesis of Z isoprenoids.
applicability to the synthesis of all conceivable stereoisomers
is yet another potential advantage of the new protocol
reported herein over the existing ones. Of various existing
methods, the Biellmann coupling and other related allyl-allyl
coupling protocols16 require several steps for homologation
of isoprenoids by one isoprene unit. Furthermore, they are
prone to stereo- and regioisomerization reactions. A recent
modification by Grigorieva17 does give the Z isomers, but
leaves much to be desired in terms of overall efficiency and
stereoselectivity, which has been <90%. Another noteworthy
method is a two-pot “H-to-T” carbometalation-cross cou-
pling tandem process introduced by us in 19806 and used in
a number of natural product syntheses.18 As efficient and
selective as it is, it still requires two steps for homologation
To demonstrate the generality of the present protocol with
respect to the E-Z geometry, two less accessible stereoiso-
mers of farnesols, i.e., (2E, 6Z)-farnesol (7) and (2Z, 6Z)-
farnesol (8), as well as (2E, 6Z, 10E)-geranylgeraniol (9)
were synthesized by using (5Z)-8-iodo-6-methyl-1-trimeth-
ylsilyl-5-octen-yne (10), prepared in 84% yield from 1b and
TMS-protected homopropargylzinc bromide under the condi-
tions B shown in Scheme 2. The results of the synthesis of
two isomers of farnesol, i.e., 7 and 8, are shown in Scheme
4.
Although highly satisfactory, the Pd-catalyzed cross-
coupling reaction of 1b proved to be distinctly more
sluggish,19 requiring 12 h at 23 °C for completion, whereas
the corresponding reaction of 1a was over in 1-2 h.
Lithiation of 10 with t-BuLi at -78 °C (45 min) followed
by zincation with ZnBr2 at -78 to 23 °C was essentially
quantitative, as judged by the yield of 8-deuterio derivative
of 10 obtained via deuteriolysis with DCl-D2O (>95% D).
It should, however, be noted that the lithio derivative 12 is
considerably less stable than its E isomer. Thus, warming
the lithio derivative of 10 to 23 °C over 1 h followed by
Scheme 3
(16) (a) Biellmann, J. F.; Ducep, J. B. Tetrahedron Lett. 1969, 3707. (b)
Altman, L. J.; Ash, L.; Marson, S. Synthesis 1974, 129. (c) Yanagisawa,
A.; Habaue, S.; Yamamoto, H. J. Am. Chem. Soc. 1991, 113, 8955. (d)
Corey, E. J.; Shieh, W.-C. Tetrahedron Lett. 1992, 33, 6435.
(17) Grigorieva, N. Ya.; Pinsker, O. A.; Moiseenkov, A. M. MendeleeV
Commun. 1994, 129.
(18) For some of the earlier applications, see (a) Whitesell, J. K.; Fisher,
M.; Jardine, P. D. S. J. Org. Chem. 1983, 48, 1556. (b) McMurry, J. E.;
Bosch, G. K. J. Org. Chem. 1987, 52, 4885. (c) Asao, K.; Iio, H.;
Tokoroyama, T. Tetrahedron Lett. 1989, 30, 6401.
(19) Although still unclear, a chelation effect due to the δ-I atom in a
putative oxidation addition product may be responsible for the observed
difference.
Org. Lett., Vol. 4, No. 2, 2002
263

98% Z 11 in 78% yield. Conversion of 11 into 7 and 8 by
using carboalumination-hydroxymethylation11 and Sato’s
alkyne hydroxmethylation-Ti-catalyzed hydromagnesiation-
methylation12 proceeded in 71 and 70% yields, respectively.
The stereoisomeric purity in each case was 98% by 13C
NMR. Although not performed, the use of 1a in place of 1b
would produce (2E,6E)- and (2Z,6E)-farnesols.
Scheme 4
Finally, the feasibility of selectively incorporating both E
and Z isomers of 1 was demonstrated by the synthesis of
>98% isomerically pure (2E, 6Z, 10E)-geranylgeranoil (9)
from 1a, 1b, 3, and 1-iodo-2-methylpropene in 4 linear steps
in 33% overall yield (Scheme 6).
Scheme 6
deuteriolysis gave a mixture of 13 and 14 in 48 and 37%
yields, respectively, with no starting lithio derivative remain-
ing (Scheme 5). Evidently, the 8-lithio derivative have
Scheme 5
Acknowledgment. We thank the National Science Foun-
dation (CHE-0080795), the National Institutes of Health (GM
36792), and Purdue University for support of this research.
We also thank Albemarle, Boulder Chemical, and Johnson
Matthey for assistance in the procurement of Al, Zr, and Pd
compounds, respectively.
undergone an intramolecular metallotropy. Despite these
potential complications, the desired homoallylzinc bromide
can be generated in nearly quantitative yield and cross
coupled with 1-iodo-2-methylpropene in the presence of
tris(2-furyl)phosphine (TFP) to give, after desilylation,
Supporting Information Available: Experimental pro-
cedures, spectral data, and spectra. This material is available
free of charge via the Internet at http://pubs.acs.org.
OL010263D
264
Org. Lett., Vol. 4, No. 2, 2002